Distributed anomaly detection algorithm based on joint classifier

Cloud computing resources access modes as well as its collaborative computing characteristics are conducive to intrusion, and it makes cloud computing security issues more worrying, especially in largescale cloud computing environment. In contrast to previous methods, our approach introduces a cloud computing intrusion detection system based on joint classifier, and makes improvement to the security of cloud computing platform. There are three main aspects. 1) Traffic flow monitoring method, 2) Flow feature selection technique, 3) Joint classification algorithm. Furthermore, our approach has two important features. First, we use feature selection technique to find out all possible combinations of the attributes from the data set, and pick out a set of attributes with the best prediction accuracy. Second, this multi-level intrusion detection method combines with K-means and Bayesian Belief Network, and it can comprehensive cascade defects of single algorithm, and make all the indicators of results more balance, to achieve an ideal result generally. According to our evaluation result, our approach can improve the detection efficiency and real-time monitor in cloud computing platform. And our multi-level algorithm is better than the single K-means and Bayesian Belief algorithm, especially in the precision rate and the overall accuracy. ©, 2015, Journal of Computational Information Systems. All right reserved.